In the frame of the lift, the bolt is at rest initially, then accelerates at 10ms-2 over 3m until it hits the floor.
From the relation v2=u2 + 2as the velocity of the bolt just before impact is sqrt(2*g*3m) = sqrt(60) = 7.74596669, and its KE is 9J. All of the KE is converted into heat energy.
The result has been calculated in the frame of the lift, so it will be the same regardless of how the lift appears to be moving to observers in other reference frames (assuming the speeds involved are sufficiently low that relativistic effects can be ignored).
You might have been confused by the fact that from some other reference frame in which the lift is moving, the gain in KE appears to be different. To take the example given in the question, if the lift is moving at 7m/s, then to a stationary observer the bolt appears to accelerate from 7m/s to 14.74596669m/s. Its KE as measured in that reference frame changes from 7.35J at the start to c32.6J just before impact, an increase of 25.26J, which is greater than the KE it gained when measured from the rest frame of the lift, and you might wonder why all of that increase KE is not converted into heat during the impact. The answer is that in the rest frame of the lift the lift does not recoil from the collision, so the lift gains no KE from the collision. However, in any other frame the recoil of the lift needs to be taken into account. When that is done, you find that some of the KE lost by the bolt is transferred into KE of the lift- the amount available to convert into heat remains as it was in the lift's rest frame.
You will see some answers which say that the bolt starts at one velocity and ends at the same velocity, so it gains no KE, which means that all of the PE lost is converted into heat. That answer is misleading- the bolt gains KE continually as it falls, and it is the KE gained by the bolt which is converted into heat at the moment of impact.